Skip to main content


Figure 3 | Breast Cancer Research

Figure 3

From: A robust classifier of high predictive value to identify good prognosis patients in ER-negative breast cancer

Figure 3

Kaplan-Meier curves for MDAhet classifier. Kaplan-Meier survival curves for the three subclasses 'good-down' (light green), 'good-up' (dark green), 'poor-down' (blue), as predicted by the Heterogeneous Mixture Discriminant Analysis (MDAhet) classifier, in the training and combined test cohorts. The class 'poor-up' is not shown due to small sample size (Table 2). Hazard ratios (HR), 95% confience intervals (CI) and log-rank test p-values are given for the predicted 'good-up' class relative to the predicted poor prognostic classes, as given by a stratified Cox-regression model with strata defined by cohorts. The Kaplan-Meier curves for each subclass is shown separately for disease-specific survival (solid lines) and distant metastasis (broken lines).

Back to article page